Hi
I was wondering how carrier scattering varies with the defects in thin film semiconductors. For example if iii-V semiconductor grows on silicon substrate because of the lattice mismatch there will be lot of misfit dislocations and growth twins. Is there any straight forward answer where it can be said that the carrier scattering increases/decreases with line defects or stacking faults. Or if I pose the question like this: how the mobility is affected by these defects. Will it increase or decrease by these defects? Or is it complicated to say? There exist lot of literature which mostly are misleading or I did not find very good answer in literature which can straight way gun down the question. Some literature says carrier scattering will be more with twin length some says it would decrease with coherent twin boundary length as twin boundaries are low scattering centers. I am not sure how to interpret. In my case I got twin formation between film and the substrate which drastically increases the conductivity. I am not sure whether to interpret that twin boundaries played a dominant role to reduce the carrier scattering and thus enhanced the mobility and conductivity or it has no effect on carrier mobility or scattering. Your thoughts will be highly appreciable.
An increased conductivity, expressed by the specific conductivity sigma, can be interpreted by the basic equation sigma = n e µ, with the density n of the charge carriers, each carrying one elementary charge e (otherwise one just has to include the respective multiplicative factor), and their mobility µ. Since usually neither n nor e vary much for a given material, an observed change in conductivity can be traced back to a corresponding change in mobility.
Hi Prof. Wagner
Thanks for your reply. But I guess your answer is not in resonance with my question. My doubt is how line defects and stacking faults/twins affect carrier mobility. ?
Thanks
Ankur
Generally any defect reduces mobility because it disturbs the spatially uniform density of states within which carriers normally travel (conduction band, valence band).
However, it is possible that the valence and conduction band have so few carriers that a high defect density in the band gap can open up a defect conduction channel that exhibits higher conductivity (but not higher mobility) than the normal extended state transport.
Ankurda,
Usually in nitride semiconductors heterojunctions there are various factors responsible for scattering mechanism,e.g.:acoustic phonon, optical phonon, ionized impurity, interface roughness, surface donor and most importantly threading dislocations. Under different ranges of temperatures one or other factor becomes more dominant than the other. In case of nitrides the dislocations are found to be more benign than other semiconductors. Generally , if the distance between two neighboring dislocations are longer than the scattering length of the electrons, no more further reduction of dislocations are of any help to enhance the mobility of the 2DEG formed at the heterojunction of AlGaN/GaN. Under that situation other factors need to taken care of.
Thanks
Abheek
Abheek
Your answer is wonderful and to the point. However, one point if you can elaborate in case if you have come across such literature or experienced of, did you observe any discrepancy in conductivity when film formed through twining. Or do you think similar explanation valid also for twins and twin boundaries?
One more question
Is there any method to calculate scattering length of electron in 2DEG? I am not aware of. Or is it case specific and complicated? If it is possible then twin boundary width and scattering length can be compared to corroborate the conductivity.
Thanks a lot for your very good input.
Hi Manuel
Thanks for your answer and your insightful comments. I do agree that carrier concentration increases because of high defect density. But I am not sure whether it is good to comment mobility has no effect with defect density. Mobility does depend with effective mass. If semiconductors contains twins, grain boundary defects then effective mass of the carrier is influenced and thus mobility. However, it ,may not be always reduced. It may increase case to case basis. I am attaching one theoretical paper on GaAs nano wire where it shows mobilty directly proportional with twin boundary length (Eq.4). Means it increases with increasing twin boundary. But on the other hand they also showed some of the cases where mobility decreases with increasing twin boundary length. Some of the other literature also told similar things. Therefore it is quite confusing to me. Let me know your views, I may be completely wrong in explanation or understanding, correct me if it is so.
Thanks
Ankur
I am not well-versed in field of the solid state physics that can throw some light on the determination of the mean free path of electrons in bulk semiconductor or the 2DEG formed at the heterojunction. See the link below that determines the carrier transport property in 2DEG formed at the heterointerface of AlGaN/GaN system.
http://scitation.aip.org/content/aip/journal/jap/82/6/10.1063/1.366137
Thanks
Abheek
Ankur and colleagues,
Good discussion on the effect of crystallographic defects on mobility of charge carriers. Basically any departure from the periodic crystallographic structure causes scattering of the moving charges and therefore impede their motion causing the decreases of their mobility compared to defect free case.
Concerning the surface plane defects such as that of twinning or the grain boundaries the orientation of the defect to the motion of the charge carriers greatly affects their mobility. It is so if the defect is parallel to the the motion of the charge carriers it will affect the volume mobility much less than if they are lying normal to the motion of the carriers.
This matter of fact has been exploited in the semi crystalline poly crystalline silicon for solar cells. Their columnar structure made the effect of the grain boundaries much less than the perpendicular grain boundary. The surface scattering of the grain boundary will be smaller as the grain size increases.
I think this may explain your findings. In case of miss matched interface the disturber layer thickness will be larger than the case of twinning.
Best wishes
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science
- Material Characterization